bug#41099: 28.0.50; TRAMP process-file ignores exit status of remote process

[Dmitry Gutov]

On 16.05.2020 15:19, Michael Albinus wrote:

> > On 14.05.2020 18:49, Michael Albinus wrote:
> > > I understand (and sympathize) both positions. However, Tramp has
> > > returned for decades no strings for process-file, so I don't expect any
> > > code in the wild which expects this.
> >
> > But is there code in the wild that expects the _current_ behavior?
>
> Don't know. But at least Philipp has reported this inconsistency, so
> there are prople who care.

Care for the remote case to behave like the local one, right? Not the 
reverse?

> > It sounds rather odd to me, given that such code would only be
> > intended to run on remote systems, but never on the local one. Is that
> > about right?
>
> Emacs has no problem to detect, whether a local process has been
> interrupted by a signal. It does it on C level.

OK, so if I understand you right, Tramp ends up doing some extra 
computations to get that info, and that makes it slower. I suppose this 
could be a reason to make the "correct" behavior disabled by default.